Our observations showed upwelling of cold, O2-depleted and nutrient-rich intermediate water in progress in the narrow inner shelf region of the transect between 78.3°W and 79°W (station 8-18), providing optimal growth conditions for phytoplankton. However, observed maximum concentrations of Chl a were low compared to common values for this extremely productive area (Chavez, 1995; Pennington et al., 2006). A large discrepancy existed between nutrient availability and particulate organic matter concentrations in the surface layer, with PON and POP levels up to an order of magnitude lower than inorganic NO3
and PO4
3-. Mean surface concentrations of NO3- and PO43- in the tropical South East Pacific for January 2009 approximated by World Ocean Atlas data (Garcia et al., 2010b) show that the waters along 10°S were in the center of a massive upwelling plume, originating from the Peruvian shelf (Fig.
1.14A and 1.14B). Repressed uptake of these upwelled nutrients in surface waters enabled the expansion of nutrient plumes far into the oligotrophic Pacific open ocean to 115°W.
Figure 1.14. Mean surface concentrations of (A) NO3
and (B) PO4
in the tropical South East Pacific for January 2009 (from World Ocean Atlas 2009; Garcia et al., 2010).
These high-nutrient, low-chlorophyll (HNLC) conditions have been reported previously for the Peruvian upwelling system (i.e. Thomas, 1979; Strickland et al., 1969; Minas and Minas, 1992).
Iron limitation of phytoplankton, especially of diatoms, has been discussed as a possible explanation (Bruland et al., 2005). However, measured iron concentrations were not limiting
along the transect (C. Schlosser, personal communication). Very high ratios of NO3- to Si(OH)4
of over 10:1 in the surface layer west of the inner shelf indicate that dissolved silica is ultimately limiting the build-up of diatom biomass (Conley and Malone, 1992; Dugdale et al., 1995). In fact, surface Si(OH)4 concentrations of <2 µmol l-1 west of the inner shelf are likely to cause Si(OH)4 limitation stress for diatoms growing under NO3- replete conditions. But if blooming of diatoms was impeded by Si(OH)4 limitation, other non-silicifying phytoplankton could have taken over.
The fact that this was not observed points to herbivorous grazing by meso- and microzooplankton controlling phytoplankton standing stocks (Minas et al., 1986; Cullen et al., 1992). Most copepods avoid O2-depleted intermediate water layers and are forced to remain in the ventilated upper part of the water column (Boyd and Cowles, 1980). The resulting
‘concentrated’ grazing pressure may have contributed to keeping microalgal biomass low. High abundance of the chlorophyll derivative phaeophorbide in surface waters of the shelf area supports this assumption. Grazing on phytoplankton by herbivores produces this phaeopigment, which serves thus as a suitable indicator for zooplankton feeding activity. Complete absence of phaeophorbide at station 21 may imply that the NH4+ maximum at this station was mainly a product of bacterial decomposition rather than of zooplankton excretion (Smith and Whitledge, 1977).
4.2. Non-‘Redfieldian’ primary production
Dominance of phytoplankton communities by diatoms is a common characteristic of coastal upwelling systems and large phytoplankton generally prosper in nutrient-rich waters. In particular diatoms can take advantage of nutrient replete conditions through high levels of maximum specific uptake rates and a quick metabolic response after vertical intrusions of nutrient-rich water (Fawcett and Ward, 2011). As already pointed out these ‘bloomers’ are adapted to exponential growth with fast cell division and accordingly their metabolism requires the synthesis of large amounts of RNA, which is generally low in N:P. Nutrient requirements based on the specific growth strategy of the phytoplankton are therefore responsible for its cellular N:P composition (Arrigo, 2005). Particulate organic matter in the water column above the inner shelf had rather low N:P ratios <16:1, consistent with increased abundances of large diatoms, dinoflagellates and cryptophytes, which are all originating phylogenetically from the red plastid superfamily featuring low cellular N:P quotas (Quigg et al., 2003; Falkowski et al., 2004).
According to Mills and Arrigo (2010) low N:P uptake of exponentially growing phytoplankton in eutrophic systems may be the main reason for a reduced availability of P* for diazotrophic phytoplankton. This point is however addressed in more detail in the next section.
Prasinophytes and chlorophytes, both belonging to the green plastid superfamily, plus the unicellular cyanobacterium Synechococcus, also contributed to primary production in the
surface water layer of the inner shelf. Their cells commonly exhibit a relatively high N:P composition >20:1 (Quigg et al., 2003; Bertilsson et al., 2003), implying that N:P ratios of the inner shelf waters would have been even lower without the occurrence of this generally smaller phytoplankton.
Nutrient conditions for the phytoplankton community changed between the inner shelf and the steep shelf break due to a drastic shift in the hydrographic structure from intense coastal upwelling to the formation of a pronounced pycnocline. Accumulation of NH4+
in subsurface layers associated with strong density gradients is a common phenomenon in well-stratified waters, where a sufficient source of organic material is available from sinking organic matter
‘trapped’ within the pycnocline (Fogg, 1991; Holmedal and Utnes, 2006). The resulting long residence time of organic particles within this layer is generating optimal conditions for regeneration by zooplankton grazing and/or heterotrophic bacteria (Saino et al., 1983;
Brzezinski, 1988). But since the absence of phaeophorbide at station 21 indicate a minor relevance of herbivorous grazing, this nutrient recycling system was presumably sustained by microbial degradation.
Even though NO3- and PO43- were present at non-limiting concentrations in off-shelf waters, biomass of large blooming PFTs was significantly lower compared to the near-shore surface waters. The phytoplankton community west of the inner shelf consisted to a large part of nano- and picophytoplankton, which is characteristic for systems of regenerative primary production (Malone, 1980). As discussed in Section 4.1, diatom growth was likely inhibited by Si(OH)4
limitation outside the center of upwelling. Size-selective grazing by mesozooplankton on large microalgae and microzooplankton may have suppressed build-up of the microphytoplankton community and relieved grazing pressure on the nano- and picoplankton communities (Richardson et al., 2004).
According to its marker pigment concentration the pico-cyanobacterium Synechococcus was the most abundant phytoplankton in the NH4+-enriched subsurface layer. With its small cell size and a metabolism adapted to low nutrient concentrations, Synechococcus is a typical representative of the ‘survivalists’. Cellular N:P is high owing to large amounts of proteins for nutrient uptake which permit maintaining net growth even under low nutrient availability (Bertilsson et al., 2003).
Chlorophytes, as part of the green plastid superfamily, are also characterized by a high N:P quota exceeding by far Redfield proportions. In the shelf slope area this group co-occurred with the abundant haptophytes and low numbers of microphytoplanktonic species, both characterized by low N:P ratios. Thus, the co-occurrence of ‘bloomers’ and ‘survivalists’ in this transition zone between shelf and oceanic waters resulted in the observed intermediate N:P ratios and can be seen as an example for the multi-specific composition of the intermediate Redfield ratio of 16:1 (Klausmeier et al., 2004a).
The oceanic section of the 10°S transect at station 22 and 24 west of 81°W was almost exclusively inhabited by one phytoplankton species, the unicellular pico-cyanobacterium Prochlorococcus, which is known to dominate photosynthetic biomass in the oligotrophic ocean (e.g. Campbell et al., 1994; Liu et al., 1997). A low-light adapted strain was distributed along the lower part of the thermo- and oxycline. Due to its photosynthetic apparatus Prochlorococcus can absorb photons at very high efficiency even at extremely low irradiances, allowing growth in waters below the nutricline down to 150-200 m depth. This is probably an essential feature of this non-diazotrophic phytoplankton to satisfy its nutrient demands for surviving in the highly stratified oligotrophic ocean (Partensky et al., 1993; Moore et al., 1995).
The high PON:POP stoichiometry (>20:1) in the open ocean section west of 80°W indicates a pronounced N-rich nutrient-acquisition machinery of the phytoplankton cells, although nutrients in the seawater were still available in sufficient quantities. Particularly high ratios (>40:1) were observed in the surface layer at 82.5°W (station 24) and in the intermediate water body between 80 and 170 m, correlating with the occurrence of the two Prochlorococcus strains. As the dominating photoautotrophic species of the oligotrophic ocean, this picoplanktonic organism is also featuring a high N:P- containing functional machinery (Bertilsson et al., 2003), allowing generally exploitation of the impoverished nutrient pools of the stratified ocean. Replete nutrient conditions even in the off-shelf waters did apparently not initiate a shift in allocation of cellular resources towards production of growth machinery in Prochlorococcus. In the course of ocean warming and the associated strengthening of water column stratification, oligotrophic regions are likely to expand and will promote growth of this high N:P assimilating picoplankton (Irwin et al., 2009). Increased uptake of N compared to P by expanding distributions of Prochlorococcus may enhance the inorganic N-deficit and generate elevated amounts of P* which could promote N-fixation (Mills and Arrigo, 2010).
4.3. Does P* control N-fixation?
A concept introduced by Deutsch et al. (2007) hypothesizes a tight spatial coupling between processes of N loss via denitrification and N gain via N-fixation. In their model simulations of global N-fixation rates Deutsch et al. (2007) assume an N:P uptake by non-diazotrophic phytoplankton according to Redfield. Bioavailable N lost via denitrification is estimated to range between 200-300 Tg yr-1 (Codispoti, 1995; Galloway et al., 1995), generating excess PO43-
and thus favouring growth of N-fixing cyanobacteria. OMZs such as off the Peruvian coast represent particularly large sinks of inorganic N. In fact, low inorganic N:P ratios <10:1 in the vicinity of the shelf sediment indicate that a significant amount of remineralized NO3- is consumed by the microbial processes of denitrification (e.g. Codispoti and Packard, 1980) and/or DNRA (Lam et al., 2009), the latter subsequently providing NH4+ for anammox, before reaching the euphotic
zone. As a result of this microbially induced N loss and the concomitant gain of P from anoxic shelf sediments, N has the potential to run into depletion before P does (Harrison et al., 1981), generating excess amounts of P available for the autotrophic community. Especially large phytoplankton in OMZ-influenced coastal upwelling areas takes advantage of the low inorganic N:P stoichiometry and removes a large portion of the P* generated in O2-depleted intermediate shelf waters from surface shelf waters before it reaches oceanic waters (see Fig. 1.5B). This non-Redfield nutrient assimilation by non-diazotrophic phytoplankton counteracts the replenishment of the local N-deficit by decoupling the microbial processes of N loss and N gain in OMZ-influenced waters.
Despite the fact that positive values of P* were present throughout the entire transect, based on the phytoplankton pigment analysis we could not detect photoautotrophic diazotrophic cyanobacteria at any of the stations sampled for PFT abundance. Trichodesmium and Crocosphaera, representing presumably the most dominant species of nitrogen-fixing cyanobacteria in the tropical ocean, both contain the carotenoid myxoxanthophyll (Carpenter et al., 1993; Mohamed et al., 2005), which was not detected at any of the stations along 10°S.
Molecular data collected during the same cruise detected genes expressing the N-fixation catalyzing enzyme nitrogenase concentrated within the OMZ on the 10°S transect (Löscher et al., unpublished results). These N-fixers were identified primarily as new clusters of probably heterotrophic bacteria. The recently discovered unicellular ‘Group A’ cyanobacteria is reported to lack photosystem II as well as photosynthetic accessory pigments (Zehr et al., 2008).
Pigment fingerprinting can consequently not be applied for identification of this exceptional group of diazotrophs. However, Löscher et al. (unpublished results) did not detect any gene copies by representatives of ‘Group A’ along 10°S. Considering this finding and on the basis of the distribution of phytoplankton marker pigments along 10°S, we conclude that no phototrophic N-fixers occurred along the transect. This implies that the ubiquitous presence of excess concentrations of P in the upper 200 m along the transect did not result in growth of N-fixing cyanobacteria and conflicts with the hypothesis by Deutsch et al. (2007). Apparently it is not possible to deduce the distribution of diazotrophic phytoplankton solely from the N:P stoichiometry of dissolved nutrients without considering non-Redfield uptake of phytoplankton.
Further crucial factors such as N availability, iron supply (Berman-Frank et al., 2001) and temperature distribution (reviewed by Stal, 2009) also have to be taken into account.
In fact, NO3- concentrations hardly dropped below 5 µmol l-1 even in the surface layer along 10°S. An ecological niche for N-fixers to enrich these waters with further N was thus not really given. Even though the low N:P supply ratio would have favoured growth of diazotrophic cyanobacteria, relatively high concentrations of NO3
(and also NH4+
) in the off-shelf surface waters may have repressed their expansion.
As already mentioned, phytoplankton growth along 10°S was not limited by the micronutrient iron (C. Schlosser, personal communication). Yet, several studies reported on a tight correlation between water temperature and the distribution of N-fixing cyanobacteria (e.g. Falcón et al., 2005; Staal et al., 2007). In particular Trichodesmium, but also unicellular diazotrophs common in warmer waters appear to have a narrow temperature range for growth and N-fixation.
Regions in the tropical and subtropical ocean characterized by water temperatures below 25°C were in most of the cases devoid of diazotrophic cyanobacteria, even with nutrient conditions (low inorganic N:P) favourable for their growth (Staal et al., 2007). Surface temperatures along 10°S never exceeded 25°C, but ranged mainly between 20-24°C in the off-shelf waters, which may have precluded the development of diazotrophic cyanobacteria despite the abundance of excess phosphate.
5. Conclusions
Our observations showed that the large oceanographic variability in the South East Pacific provides the habitat for a multitude of phytoplankton communities, forming a highly diverse photoautotrophic ecosystem from large diatoms in the near-shore upwelling areas to small picoplankton in the open ocean. Total phytoplankton biomass was kept low probably by zooplankton grazing in combination with an offshelf Si(OH)4 limitation of diatoms. Horizontal as well as vertical gradients in hydrography and nutrient distribution have a crucial impact on the taxonomical composition of the phytoplankton community by selecting for different types of growth strategies. Associated differences in specific nutrient requirements caused strong deviations in biomass elemental composition from the Redfield ratio, emphasizing
non-‘Redfieldian’ nutrient assimilation by phytoplankton as a major driver in ecological stoichiometry.
In order to evaluate the role of P* as a control for the abundance of N-fixing cyanobacteria in the eastern boundary current areas further field data, including especially rates of N-fixation, are necessary. Even though P* serves as the primary driver promoting growth of diazotrophic cyanobacteria on a global scale, its presence is a necessary but not sufficient condition for its development locally. Additional factors influencing the distribution of N-fixers on a local scale, such as N availability, iron supply and temperature, have to be considered.
Acknowledgments
We gratefully acknowledge technical assistance and support by Peter Fritsche, the CTD-Team and the Crew of R/V Meteor on-board the M77/3 cruise. Thanks also to Kai G. Schulz for help preparing the section plots. This work is a contribution of the Sonderforschungsbereich 754
"Climate-Biogeochemistry Interactions in the Tropical Ocean" (www.sfb754.de) which is supported by the German Science Foundation (DFG).